Condition Assessment of Civil Structures under Earthquake Excitation

This paper presents a nondestructive rapid condition or health assessment procedure of three-dimensional civil structures following strong earthquakes. It is a nonlinear time domain system identification procedure and structures are represented by finite elements. The excitation time history may not be available soon after an earthquake and is not required to implement the procedure. The response time histories measured at a small part of a large structure, denoted as a substructure, can be noise-contaminated. The location and severity of defects in an element are evaluated by comparing its stiffness parameter with the previous value or deviation from the other similar members. The basic unscented Kalman filter concept available in the literature is significantly improved, increasing its implementation potential. The improved concept is developed by integrating the iterative least-squares, weighted global iteration, and objective function algorithms. The proposed concept is verified by assessing the health of large three-dimensional structures excited by earthquake time histories. The procedure is capable of identifying defects with different levels of severity in small or large civil structures. This study also indicates that the inspection outcomes are not sensitive to the location and configuration of the substructure. The capability and efficiency of the procedure are demonstrated with three informative examples.